Step 4 sounds risky. What happens when hive goes through a big bear market and HBD exceeds the 10% debt ratio? HBD will constantly be valued under a dollar, because it can't be converted to 1$ worth of hive. In that case the HIGH and LOW limits will have to be constantly readjusted so that the current HBD market price is centered between both numbers.
Otherwise, the DHF can run into big losses quickly. For instance, Hive's marketcap is 50 million USD. But, HBD's supply is 10 million. That means that one HBD now converts to only $0.5 worth of hive. If LOW remains at $0.99, while HBD trades at $0.6, then the DHF will be executing step 4 while it should execute step 3.
This brings up the question, who sets HIGH and LOW? If it's hardcoded into the proposal, it's going to run into the problem I mentioned, and we will need to write a new proposal and vote it in quickly, which is impractical.
Another way is to hardcode HIGH and LOW into the proposal, and then have the witnesses provide purposely the wrong price feed for Hive's price so that the blockchain thinks HBD remains at 1$. But this method will affect other aspects like posts payouts, so also not practical.
A way to make it work would be to have witnesses set HIGH and LOW similar to how they set hive's price feed.
Let me know if I missed something or if this is the way you thought it would work.
Another idea I would prefer is the following, but it is simply a matter of choice, not mechanics, and I would support the proposal even if it is ignored.
I would prefer if the profits or extra coins resulting from the arbitrage would be burned instead of sent back to the DAO. That would benefit everyone on hive, including the DAO since higher price for hive would mean a larger budget.
How it would work:
@null. So if HBD is trading at $1.50, $0.50 worth of hive will be burned for every printed HBD and $1 worth of hive will be sent to the fund.
RE: Request for comments: HBD stabilization DHF proposal